TOWN AND COUNTRY: ROMAN BATH AND ITS HINTERLAND Peter Davenport Roman Bath, Aquae Sulis, was not a city in any legal sense recognised by a Roman civil servant, but it was a wealthy and bustling town, with an architectural grandeur to rival any city in the Province of Britannia. Some of this wealth came direct from the touristic and medico-ritual exploitation of the hot springs. Much came from the services that a local, we might today say market, town provides for its agricultural and industrial hinterland. Bath was a market centre, a route node, and may have played some sort of administrative role, the precise nature of which will always be a subject of speculation, and which could, anyway, have changed over the centuries of Roman rule. In this article we shall look, as far as the evidence allows, for the pattern and extent of that hinterland, and suggest a little of how the one interacted with the other. Defining the hinterland is a subjective process, but it can be based on various consider­ ations. One should look at available resources, the communications network, competing centres, and the distribution of rural settlement itself. We must also glance upwards at the settlement hierarchy to notice that the town itself is part of the hinterland of other cities. Economically, Bath probably looked to Cirencester; administratively, to Winchester. For the purposes of this discussion, however, we shall define our region in a simple way. In an era where the maximum economic speed of long­ distance travel was walking pace, a locally cohesive area would not be bigger than could be crossed in less than a day. Therefore, no small town could have a hinterland more than a day's return journey in radius- say 15 to 20km.1 Ignoring the special status due to it as a spa and religious centre, Bath will be seen to fit well into this pattern. It might seem obvious to start from a consideration of the transport system. However, the routes of the main roads around Bath are subject to much uncertainty,2 and the Avon may not have been navigable in the Roman period.3 The local communication network, particularly relevant for our purposes, is almost entirely unknown. The known and likely roads are shown on the maps 4 (Fig. 1). Nonetheless, the position of Aquae Sulis at a route node is important in understanding its relationship to its C1:J >-;:) ~ t:rl ::00 CJ ~ t:rl aq Fig. 1 The pattern of settlement and the topography around Aquae Sulis. Squares represent villas (larger squares, the larger, more luxurious ones), and triangles, other farms. The hatched areas are nucleated settlement. Contours are at SOm intervals, and land below SO mOD is shaded. Roads are shown as thick lines, probable local tracks and paths as thinner ones. The town is in the centre, with its late walls shown by a thick line. ROMAN BATH AND ITS HINTERLAND 9 local region. It sits by the crossing point of the Fosse Way over the Avon, and that road's junction with the roads from London, Poole harbour, Cirencester via Lansdown, and Sea Mills (for the Severn crossing). It is also on the Avon valley route through the Cotswolds. Even locally the traveller is almost compelled by the topography to use the Avon valley eventually. Little traffic would not pass through, or near, the town. Missing from this known picture are the local roads and paths linking settlement, town and villa. Prehistoric trackways and footpaths such as Hollies Lane across Charmy Down can be identified, and others can be suggested. A number of present-day lanes and rights-of-way seem to link Roman sites and can be seen from the map (Fig. 1). These tracks, if correctly identified, imply regular contact between the components of the pattern. We must now look at these components: other towns in the area, Aquae Sulis itself, villages, farms and villas. Local Towns and Aquae Sulis Bath's urban neighbours can indicate the likely broad limits of its hinterland. The position of nearby small towns such as Verlucio (Sandy Lane, near Lacock), Abonae (Sea Mills, near Bristol), and sites such as Camerton and Nettleton, probably confirm Bath's local area as being around 15-20km in radius. They existed as secondary market centres on the borders of the regions of the larger towns. Sea Mills was the port of the region, particularly important to the south Wales garrison. It was occupied early and probably always had a military presence. The site is not well known, but covered a fairly extensive area, and buildings, burials and finds indicate a substantial small town.5 Camerton had a conquest-period fort on the site of a high-status native site.6 It was little more than a substantial village, but the later industrial specialization in pewter manufacture, and its position on the main road, arguably made it a small town.? Verlucio is unexcavated, but its status as an imperial posting station, and the material recovered from it, support the claim that it, too, was a small town. Nettleton Shrub8 too was small, but intensively developed, with considerable "public works" in the form of river canalizations, stone revetting of hillsides, streets and public buildings and provision of an elaborate temple to Apollo. The temple and the trade it attracted (made possible by the Fosse Way), plus pewter manufacture, were the reasons for its wealth. 10 PETER DAVENPORT It is possible that Aquae Sulis served a wider zone owing to its great distance from the nearest major administrative and market centres. The possibility that Bath was situated in a region under direct imperial control for some of the Roman period (see below) might have encouraged its development as a centre of services for a zone that was separated from the normal civilian administration, and fitted it for an administrative role when the area was demilitarized. However, the existence of the other local small towns, and also Shepton Mallet and Mildenhall, indicates that one centre alone could not deal with all matters. There must have been some hierarchy and presumably interpenetration of spheres of influence, but these towns conveniently limit our area of study. At Aquae Sulis two early foci have been recognised: one around the hot springs; the other around a crossing point of the Avon, just south of Cleveland Bridge (Fig. 2). The nature of the earliest occupation around the springs is unclear. A metalled road existed to the west of the spring, possibly as early as c.SO AD or later in the first century, and at the same time a drain was laid parallel to it a little way south. The baths and temple were erected in the 60s or 70s AD/ but it seems likely that, at first, the area immediately around the great monuments was not intensively occupied at all. There is evidence of timber buildings, streets or yards emerging from recent work, dating to the first and early second centuries, but these may not be strictly urban in nature, merely related to the activities of those running the baths and temple. As at some similar sites in Gaul the complex may well have stood in the countryside at the beginning of the period, perhaps originally as an army facility. From the later second century onwards we see many buildings with hypocausts and mosaics, probably houses. There was an infilling of open spaces, and a replanning of this central part of the town, with shops and industry appearing. This can all be ascribed to the spa's growth in popularity and wealth. The peak of this growth in the late third and fourth centuries can be related to changes in the countryside described below. At Walcot, where the Fosse Way joined a newly discovered road leading to the River Avon and Bathwick (destined for Poole harbour), substantial evidence survives of settlement from the earliest Roman period (Fig. 3). The quantity and quality of finds imply a military presence nearby. The flat, but raised and well-drained gravel "island" across the river around StJohn's church suggests that Bathwick may well be the site of the long-sought Roman fort, and would provide a context for the growth of the settlement and a source for the objects found there.10 This may also explain why the road system as we know it focuses on this ROMAN BATH AND ITS HINTERLAND 11 '\_ I ' \ \ \ \ '. Fig. 2 The two foci of Aquae Sulis, superimposed on the street plan of modern Bath. The walled area is shown, the roads (thick solid and dashed line) and the areas of settlement outside the walls. ...... N 'U ~:;;;:! Ci ~ ~ 'U q0 Fig. 3 The Walcot focus. The broken horizontal shading indicates the confirmed spread of intense occupation. Stray finds indicate a much wider area. Based on the 1992 Ordnance Survey 1:1250 map with the permission of the Controller of Her Majesty's Stationery Office.© Crown Copyright. ROMAN BATH AND ITS HINTERLAND 13 area, not the springs (Fig. 1). It is not just that they meet here, but that they are clearly aligned on this point from some distance away. As in the centre, there is a phase of replanning, perhaps in the late second or third century, but with an increased sense of bustle and activity. While trade and industry seem more important than in the centre, we also find clear evidence of high­ quality, probably domestic, architecture. By the third and fourth centuries the two foci had fused, indicating an urban area of about sixty acres. The Countryside (i): Farms and Villages Agriculture was the basis of the Roman economy.